The overlapping RNA-binding domains of p33 and p92 replicase proteins are essential for tombusvirus replication

The overlapping RNA-binding domains of p33 and p92 replicase proteins are essential for tombusvirus replication

Available online at www.sciencedirect.com R Virology 308 (2003) 191–205 www.elsevier.com/locate/yviro The overlapping RNA-binding domains of p33 an...

631KB Sizes 0 Downloads 50 Views

Available online at www.sciencedirect.com R

Virology 308 (2003) 191–205

www.elsevier.com/locate/yviro

The overlapping RNA-binding domains of p33 and p92 replicase proteins are essential for tombusvirus replication Zˇivile˙ Panaviene˙, Jannine M. Baker, and Peter D. Nagy* Department of Plant Pathology, Agricultural Science Building-N, University of Kentucky, Lexington, KY 40546, USA Received 17 July 2002; returned to author for revision 4 September 2002; accepted 20 November 2002

Abstract Two of the five viral-coded proteins of tombusviruses, which are small, nonsegmented, plus-stranded RNA viruses of plants, are required for replication in infected cells. These replicase proteins, namely, p33 and p92, of cucumber necrosis virus are expressed directly from the genomic RNA via a readthrough mechanism. Their overlapping domains contain an arginine/proline-rich RNA-binding motif (termed RPR, which has the sequence RPRRRP). Site-directed mutagenesis of p33 expressed in Escherichia coli, followed by a gel shift assay, defined two of the four arginines as required for efficient RNA binding in vitro. In vivo testing of 19 RPR motif mutants revealed that the RPR motif, and therefore the ability to bind RNA, is important for the replication of tombusviruses and their associated defective interfering (DI) RNAs. Mutation within the RPR motif also affected the ratio of subgenomic versus genomic RNAs in infected cells. To test whether the RPR motif is essential for the function of either p33 or p92 in replication, we used a two-component system developed by Oster et al. (1998, J. Virol. 5845–5851), in which p92 was expressed from the genomic RNA of a tombusvirus, while p33 was expressed from a DI RNA. The protoplast experiments with the two-component system revealed that the RPR motif is essential for the replication function of both proteins. Interestingly, mutations within the RPR motif of p33 and p92 had different effects on RNA replication, suggesting different roles for the RNA-binding motifs of these proteins in tombusvirus replication. © 2003 Elsevier Science (USA). All rights reserved. Keywords: Replication; RdRp; RNA binding; Arginine-rich motif; Complementation; Defective interfering RNA; Subgenomic RNA; Cucumber necrosis virus; Protoplast

Introduction Replication of RNA viruses in infected cells is performed by the viral replicase. The replicase is a protein complex that includes the viral-coded RNA-dependent RNA polymerase (RdRp) and additional viral replicase protein(s) and host factors (Buck, 1996, 1999; Diez et al., 2000; Lai, 1998; Noueiry et al., 2000; Quadt et al., 1993). In addition, the viral genomic RNA must also be recruited to the viral replicase complex (Ahlquist, 2002; Buck, 1996, 1999). After the assembly of the replicase complex and the recruitment of the viral genomic RNA, a complementary (minus-strand) RNA is made. This is followed by plusstrand RNA synthesis utilizing the minus-strand intermedi* Corresponding author. Fax: ⫹1-859-323-1961. E-mail address: [email protected] (P.D. Nagy).

ates. The replication process can be very efficient for several viruses, including tombusviruses, which are plus-stranded, single-component RNA viruses of plants. Interestingly, tombusviruses are frequently associated with defective interfering (DI) RNAs that are derived entirely from the genomic RNA. The most frequently occurring DI RNAs contain three or four short noncontiguous segments of the genomic RNA (Fig. 1, Burgyan et al., 1991; Chang et al., 1995; Finnen and Rochon, 1993; Hillman et al., 1987; White and Morris, 1999). DI RNAs replicate efficiently with the help of the tombusvirus replicase; therefore they are useful in studying replication in vivo and in vitro (Nagy and Pogany, 2000; Panavas et al., 2002; White and Morris, 1999). Tombusviruses code for two replicase proteins, namely, p33 and p92, both of which are essential for virus replication in plants (Kollar and Burgyan, 1994; Oster et al., 1998;

0042-6822/03/$ – see front matter © 2003 Elsevier Science (USA). All rights reserved. doi:10.1016/S0042-6822(02)00132-0

192

Z. Panaviene et al. / Virology 308 (2003) 191–205

Fig. 1. The RPR motif in the CNV p33 is essential for RNA binding in vitro. (A) Schematic representation of the gCNV RNA (shown in 5⬘ to 3⬘ orientation). Among the five genes, p33 and p92, which are essential for tombusvirus replication, are directly expressed from the genomic RNA via a translational readthrough mechanism. The other three genes are the coat protein, the movement protein, and the suppressor of gene silencing (shown with black boxes) that are expressed from two subgenomic RNAs (not shown). The previously defined RPR motif in the overlapping region of p33 and p92 (Rajendran and Nagy, submitted for publication) is shown with a gray box and the actual amino acids are specified. We used the following symbols for the amino acids: P, proline; R, arginine; A, alanine; and S, serine. The deleted sequence in construct ⌬RPR is shown with dashes. (B) A representative gel shift analysis of protein/RNA complexes. The RNA probe (lane p, containing the probe only) was the 5⬘ 169-nt segment of DI-72(⫹) RNA, which was 32P-labeled using in vitro transcription. The same amount of RNA probe (2 ng) was added to each lane in the gel, while the WT and ⌬RPR proteins were added (0.5 ␮g each) to the lanes as indicated below the gel.

Scholthof et al., 1995). Both replicase proteins are expressed directly from the genomic RNA that requires ribosomal readthrough of the termination codon of the p33 ORF to generate p92 (Scholthof et al., 1995). Due to this expression strategy, the N-terminal portion of p92 overlaps with p33. p92 is predicted to function as an RdRp, while the function of p33 is currently unknown. Both p33 and p92 have been proposed to be part of the tombusvirus replication complexes (Scholthof et al., 1995). We have demonstrated recently that both p33 and p92 could bind to tombusvirus RNAs in vitro (Rajendran and Nagy, submitted for publication). Gel mobility-shift assays performed with deletion derivatives of p33 expressed in Escherichia coli revealed that an arginine/proline-rich motif (termed RPR motif), which is conserved among tombusviruses, is critical for efficient RNA binding. The corresponding region in the overlapping domain of p92 may also bind to RNA, although this has not been confirmed. This is because p92 has a second RNA-binding domain within its unique C-terminal region (Rajendran and Nagy, submitted for publication); therefore mutations within the RPR motif in p92 did not abolish RNA binding. In this report, we demonstrate that the RPR motif is essential for tombusvirus replication in vivo. Detailed site-specific mutagenesis of the RPR motif revealed that two of the four arginines within the RPR are critical for virus replication in protoplasts and for RNA binding in vitro. We also demonstrate that the RPR motif is essential for replication in both p33 and p92. This was achieved by using a two-component system, in which

p92 was expressed from the viral genomic RNA and p33 was expressed from a DI RNA (Oster et al., 1998). We found that mutations within the RPR in either p33 or p92 inhibited replication in the above two-component system. The nonreplicating DI RNAs carrying mutated p33 could be rendered replication-competent in the presence of the wildtype (WT) tombusvirus RNA, excluding the possibility that the mutations debilitated the ability of the DI RNA to replicate. Overall, these experiments supply direct evidence that a functional RPR motif in both p33 and p92 is essential for replication.

Results Essential role for the arginines of the arginine/prolin-rich motif in RNA binding in vitro The RPR motif present in both p33 and p92 of Tomato bushy stunt virus (TBSV) and Cucumber necrosis virus (CNV) (Fig. 1A) contains two prolines and four arginines that are the primary candidates to bind to the negatively charged viral RNA (the actual sequence of the RPR motif is R637P640R643R646R649P652, Rajendran and Nagy, submitted for publication). Accordingly, deletion of the entire RPR motif from the full-length p33 of CNV (see construct ⌬RPR, Fig. 1) inhibited the in vitro RNA-binding activity of ⌬RPR (expressed and purified from E. coli, see below) by ⬃80% when compared to the full-length WT p33 protein

Z. Panaviene et al. / Virology 308 (2003) 191–205

(Fig. 1C). Since the level of expression of the WT p33 is low in E. coli and it also shows poor solubility (not shown), we expressed a 30-amino-acid-long region of p33 (which was fused to the maltose-binding protein, MBP), including the RPR motif in E. coli. This 30-amino-acid-long region of p33 was demonstrated earlier to bind RNA efficiently (Fig. 2 and Rajendran and Nagy, submitted for publication). To test which amino acids within the RPR motif play an important role in RNA binding, we have generated 19 mutants of the CNV p33 that had one or more altered amino acids within the RPR motif (see Fig. 2A). One set of mutants contained one, two, or three alanine mutations replacing either arginines or prolines. The second set of mutants contained arginine to lysine mutations, which maintained the positive charge of the RPR motif. The third set of mutants had one, two, or three of the arginines deleted. The fourth set of mutants contained one or two extra arginines that increased the overall positive charge of the RPR domain. The purified truncated p33 proteins carrying 1 of the 19 mutations or the WT sequence within the RPR motif were used in similar amounts in gel mobility-shift assays in combination with the 5⬘ 169-nt segment of DI-72 RNA, a prototypical DI RNA (White and Morris, 1994). These experiments demonstrated that replacing the first proline (P640 in construct P640⫺A, lane 2, Fig. 2A) or the second proline (P652 in construct P652⫺A, lane 3, Fig. 2A) with alanines did not alter the ability of the truncated p33 to bind RNA by more than 10% when compared to the truncated p33 carrying the WT RPR motif (construct WT, Figs. 2B and C). Substitution of alanines for both prolines (P640 and P652 in construct P640 – 652⫺A, lane 4, Fig. 2A), however, decreased RNA binding by ⬃30%. In contrast, replacing either or both of the prolines with arginine(s) increased RNA binding by 2.5- to 3-fold when compared to WT (constructs P640⫺R, P652⫺R, and P640 – 652⫺R, lanes 5–7, Figs. 2A–C). Separate substitutions of each arginine with alanines in the RPR motif resulted in a 30 –50% decrease in RNA binding when compared to WT (constructs R637⫺A, R643⫺A, R646⫺A, and R649⫺A, lanes 1, 8, 10, and 12 in Fig. 2A). Combined replacement of three of the arginines (the second, third, and fourth arginines in the RPR motif, i.e., R643, R646, and R649) with alanines decreased RNA binding by 42% (construct R643– 649⫺A, lane 19 in Figs. 2A–C). Similarly, deletion of one, two, or three of the arginines reduced RNA binding significantly (by ⬃40 to 60%, see constructs ⌬R, ⌬2R, and ⌬3R, lanes 16 –18 in Figs. 2A–C). On the contrary, single lysine substitutions for the second or fourth arginines within the RPR motif (mutants R643⫺K and R649⫺K, lanes 9 and 13 in Figs. 2A–C) increased RNA binding by 32 and 280%, while a similar substitution at the third arginine (mutant R646⫺K, lane 11 in Figs. 2A–C) actually decreased RNA binding by ⬃40%. This suggests that the third arginine (R646) within the RPR motif is important for RNA binding and even the positively charged lysine is detrimental in that location. Insertion of

193

one or two arginines into the RPR motif did not increase RNA binding by more than 10 –20% (mutants R643⫹R and R643⫹2R, lanes 14 and 15 in Figs. 2A–C). Overall, these in vitro experiments suggest that the arginines in the RPR motif play important roles in RNA binding. Since the above RNA-binding studies included a 30amino-acid fragment of p33, it is possible that the RPR motif mutations (Figs. 2A–C) may have different roles and/or structures when present in the full-length p33. Therefore, we expressed three RPR motif mutants, namely, P652⫺A, R649⫺K, and R643– 649⫺A (see constructs 3, 13, and 19 in Fig. 2A), as full-length p33 proteins fused to the MBP domain (Fig. 3). Due to the difficulty of obtaining large amounts of purified full-length proteins (unlike in the case of the truncated p33), we used reduced but comparable amounts of these proteins and the WT p33 in the gel shift experiments (Fig. 3). Quantification of the bound RNA to the full-length p33 revealed that mutation P652⫺A showed activity that was comparable to that of the full-length WT p33, while R649⫺K bound 65% better and R643– 649⫺A bound only at a 45% reduced level of the WT p33 (Figs. 3A and B). Comparison of the data obtained with the RPR motif mutations present in either the full-length or the truncated p33 showed a similar trend in the overall effect on the RNA-binding activities of these RPR motif mutations (Figs. 2 and 3). Therefore, we suggest that the sequences within the RPR motif in p33 play major roles in RNA binding. The arginine/proline-rich RNA-binding motif is essential for tombusvirus replication in protoplast To test whether the RPR motif plays an essential role during tombusvirus replication, we introduced the above RPR motif mutations into the full-length genomic CNV (gCNV) RNA (see Materials and Methods). Note that due to the expression strategy of CNV, both p33 and p92 should carry the same RPR motif mutation during infection (see below). First, tombusvirus replication was tested in cis (i.e., the mutations were present on the replicating gCNV RNA) in Nicotiana benthamiana protoplasts after electroproration of the in vitro transcribed full-length gCNV RNA transcripts. Second, we also tested the ability of the RPR motif mutants of CNV to support trans replication (i.e., the replication of DI-72 RNA that must use p33/p92 in trans for its replication) by using coelectroporation of the mutated gCNV RNA transcript and the WT DI-72 RNA. It is well established that the TBSV-associated DI-72 RNA can use the replicase proteins generated by the CNV helper virus for replication (White and Morris, 1994; Oster et al., 1998). The accumulation of gCNV RNA and DI-72 RNA in protoplasts was studied using Northern blot with CNV-specific and DI RNA-specific probes, respectively. The amounts of total RNA loaded on gels were adjusted based on the host ribosomal RNA (Figs. 4 and 5A). The protoplast experiments revealed that 7 of the 19 RPR mutants supported gCNV RNA accumulation in cis and DI

194

Z. Panaviene et al. / Virology 308 (2003) 191–205

Fig. 2. Defining the amino acids within the RPR motif in CNV p33 that are essential for RNA binding in vitro. (A) The names and amino acid sequences of the 19 RPR motif mutants tested in this work are shown. The mutated amino acids are in boldface, while the deleted sequences are shown with dashes. We used the following symbols for the amino acids: A, alanine; P, proline; R, arginine; and K, lysine. (B) A representative gel shift analysis of protein/RNA complexes is shown. The RNA probe (lane p, containing the probe only) was prepared as described in the legend to Fig. 1. The same amount of RNA probe (2 ng) was added to each lane in the gel, while the proteins (1 ␮g each) were different, as indicated below the gel. Lane mb indicates the sample containing the maltose-binding protein, while lane WT depicts the truncated p33 with the WT RPR motif, which was expressed and purified from Escherichia coli as a 30-amino-acid fragment of p33 fused to MBP. Lanes 1–19 depict the truncated p33 with the RPR motif in the order as shown in A. The migration of the free probe and the RNA/protein complex is indicated on the left. Note that the migration of the RNA/protein complex can be different in the case of particular p33 mutants when compared to WT, probably due to the changes in the charge of the protein. (C) The relative RNA-binding efficiencies of the given truncated p33 mutants are shown compared to WT (100%) from three separate experiments. The amounts of free probes in each lane were measured using a phosphorimager. Light gray bars represent those mutants that showed decreased efficiencies in RNA binding.

Z. Panaviene et al. / Virology 308 (2003) 191–205

Fig. 3. Effects of amino acid changes within the RPR motif on RNA binding in vitro when present in the full-length p33 are shown. (A) The names of the RPR motif mutants and the corresponding amino acid changes are the same as shown in Fig. 2. Two separate representative gel shift analyses of protein/RNA complexes are shown. The RNA probe was prepared as described in the legend to Fig. 1. The same amount of RNA probe (2 ng) was added to each lane in the gels, while the proteins (0.5 ␮g each) were different, as indicated below the gels. (B) The relative RNAbinding efficiencies of the particular p33 mutants are shown in comparison with the WT (100%). The amounts of free probes in each lane were measured using a phosphorimager.

RNA replication in trans at 50 to 95% of the WT level (constructs R637⫺A, P640⫺A, P652⫺A, P640 – 652⫺A, P640⫺R, R649⫺A, and P649⫺K, lanes 1–5 and 12 and 13, Figs. 4A–C and 5). Interestingly, several of the above mutants supported trans replication of the DI RNA with 10 – 20% higher efficiency than the level of cis replication of the corresponding mutant gCNV. These data suggest that none of these four amino acids (i.e., R637, P640, R649, and P652) within the RPR motif is essential for cis or trans replication of tombusviruses in protoplast. In contrast, replacing the third arginines (R646) with alanine reduced gCNV and DI-72 RNA replication by over 90% (mutant R646⫺A, lane 10 in Figs. 4A–C and 5), suggesting that this arginine is important for tombusvirus replication. Replacement of the

195

second arginine (R643) with alanine reduced gCNV RNA replication by 80%, but affected DI RNA replication by 10% only (mutant R643⫺A, lane 8 in Figs. 4A–C and 5), suggesting that the second arginine is more important for cis replication than for trans replication. The essential role for the arginines within the RPR motif for tombusvirus replication was further supported by the observation that combined replacement of three of the arginines (the second, third, and the fourth, i.e., R643, R646, and R649, see mutant R643– 649⫺A, lane 19 in Figs. 4A–C and 5) with alanines completely abolished tombusvirus and DI RNA replication. Interestingly, replacing separately the two important arginines (the second and third arginines, mutants R643⫺K and R646⫺K, lanes 9 and 11, Figs. 4A–C and 5) with lysines decreased tombusvirus or DI RNA replication by 95 to 98%. This is somewhat surprising since both lysine and arginine are positively charged. In contrast, the lysine for arginine substitution at the fourth arginine position in the RPR motif had smaller effect on virus replication (50% of that of WT, see construct R649⫺K, lane 13, Figs. 4A–C), further indicating that this arginine may not be essential for tombusvirus and DI RNA replication. To test if the number of arginines within the RPR motif is important for tombusvirus replication, first we used insertion mutants that contained one or two extra arginines between the two prolines in the RPR motif (see mutants R643⫹R and R643⫹2R in Fig. 2A). Mutant R643⫹R with one extra arginine supported gCNV and DI-72 RNA replication only at a reduced efficiency (⬃5%, lane 14, Figs. 4A–C and 5), while addition of two extra arginines in mutant R643⫹2R abolished tombusvirus or DI RNA accumulation completely (lanes 14 and 15 in Figs. 4A–C and 5). The second set of constructs had deletion of one, two, or three arginines in the RPR motif that resulted in complete inhibition of tombusvirus replication (mutants ⌬R, ⌬2R, and ⌬3R, lanes 16 –18, Figs. 4A–C and 5). The third set of mutants carried proline to arginine mutations that did not alter the length of the RPR motif (see mutants P640⫺R, P652⫺R, and P640 – 652⫺R, Fig. 2A). Interestingly, one of the mutants supported gCNV replication at 60% of WT level, although the replication of DI-72 RNA was somewhat less efficient (20% of WT level, lane 5, Figs. 4A and B and 5A and B). The two other mutants replicated poorly in protoplast (5–10%, lanes 6 and 7, Figs. 4 and 5A and B). The data obtained with these deletion, insertion, and replacement mutants indicate that, in general, the number of arginines in the WT RPR motif is optimal for tombusvirus replication. We also tested if the relative accumulation of subgenomic (sg)RNAs (i.e., the ratio between the sg1 and sg2 RNAs versus gCNV) was different with all the above mutants in comparison with the WT in protoplast. Four mutants, including R643⫺A, R643⫺K, R646⫺K, and R643⫹R, supported the relative accumulation sg2 RNA at 4- to 20fold higher level than the WT did (Fig. 4C). The increase in the relative accumulation of sg1 was much less pronounced (between 2- and 4-fold, Fig. 4C) with these four mutants. In

196

Z. Panaviene et al. / Virology 308 (2003) 191–205

Fig. 4. Relative in vivo accumulation of gCNV RNA carrying mutations within the RPR motif. The full-length, infectious gCNV RNA (2 ␮g) carrying a given mutation within the RPR motif was electroporated into N. benthamiana protoplasts (5 ⫻ 105 cells per experiment). Total RNA was isolated, electrophoresed, and blotted onto a membrane, followed by probing with 32P-labeled RNA specific for gCNV (see Materials and Methods). The same amount of total RNA was used for loading onto the gels based on the estimation of the host ribosomal RNA in ethidium bromide-stained gels (not shown, see Fig. 6 as example). The positions of the gCNV (g), subgenomic RNA1 (sg1), and subgenomic RNA2 (sg2) are depicted on the left. The names of the constructs are shown on the top. The samples were taken 24 (top) and 48 (bottom) h after electroporation. The experiment was repeated three times. (B) The relative accumulations of gCNV RNA for the WT and the p33/p92 mutants are shown (WT was chosen as 100%). The amounts of gCNV in each lane were measured in the samples taken after 48 h of incubation using a phosphorimager. (C) The percentages of sg1 RNA (dark bar) and sg2 RNA (gray bar on the right) accumulation were calculated based on the accumulation of the corresponding gCNV RNA for the WT and the p33/p92 mutants. The data represent the samples taken after 48 h of incubation.

contrast, mutant P640 – 652⫺A generated sg2 RNA less efficiently (a drop of 40% in relative accumulation) than the WT did (lane 4 in Figs. 4A–C). The difference in the ratio of sg2/gCNV between P640 – 652⫺A and WT is significant

after both 24 and 48 h incubation and we confirmed this by repeating the experiment two times (not shown). Overall, this observation suggests that mutations within the RPR motif can affect subgenomic RNA synthesis/accumulation.

Z. Panaviene et al. / Virology 308 (2003) 191–205

197

replication at detectable levels. This was done by RT–PCR, followed by sequencing (in the vicinity of the RPR motif) of the progeny gCNV RNA obtained from protoplasts 48 h after electroporation. These experiments confirmed that the original mutations in the RPR motif were stably maintained during the course of infection (not shown). This observation suggests that the original mutated p33/p92 proteins and not putative revertants or WT contaminants were responsible for supporting tombusvirus replication in protoplast (Figs. 4A–C and 5). Essential role for the RPR motif in p33 during tombusvirus replication in protoplast

Fig. 5. Relative trans replication efficiency of DI-72 RNA by the gCNV helper carrying mutations within the RPR motif. The full-length DI-72 RNA (1 ␮g) (White and Morris, 1994) was coelectroporated with the infectious gCNV RNA (2 ␮g) carrying a given mutation within the RPR motif into N. benthamiana protoplasts (5 ⫻ 105 cells per experiment). Sample preparation and Northern blotting were done as described in the legend to Fig. 4, except the probing was done with 32P-labeled RNA specific for DI-72 RNA (see Materials and Methods). The position of the DI-72 RNA is depicted on the left. The names of the gCNV helper constructs are shown on the top. The samples were taken 24 (top) and 48 h (bottom) after electroporation. The experiment was repeated three times. (B) The relative trans replication of DI-72 RNA in the presence of the WT helper or one of the p33/p92 mutants is shown (WT was chosen as 100%). The amounts of DI-72 RNA in each lane were measured in the samples taken after 24 (left bars) and 48 (right bars) h of incubation using a phosphorimager.

It is possible that unwanted mutations are introduced during in vitro mutagenesis, in spite of the use of the high-fidelity Pfu Turbo DNA polymerase to introduce the site-directed mutations (see Materials and Methods). However, sequencing of a 500- to 600-nt region around the targeted RPR motif mutations did not reveal any mutated positions, except the target ones (not shown). Also, we used two independently generated clones for P652⫺R, R643– 649⫺A, and ⌬3R in the protoplast experiments, which in each case gave results comparable to that shown in Fig. 4A (lanes 6, 18, and 19), i.e., poor accumulation for P652⫺R, and an undetectable level of accumulation in the case of R643– 649⫺A and ⌬3R (not shown). Since it is very unlikely that the proven mutagenesis method used in this work (see Materials and Methods) would introduce the same second-site mutation for two separate clones, we conclude that the effect on the reduction of CNV RNA accumulation is due to specific targeted mutations introduced into the RPR motif. Since RNA viruses can accumulate mutations quickly during infection, we tested the stability of the mutated gCNV RNA for the 12 constructs that supported gCNV

Although the above experiments demonstrated that the RPR motif plays an essential role in tombusvirus replication, due to the overlapping domains in p33 and p92, they could not answer whether the RPR motif is important for the function(s) of p33 or p92 or both. To test this, we used a two-component, complementation-based tombusvirus replication system developed by Oster et al. (1998). Briefly, in the two-component system, the p92 is expressed from the gCNV RNA that carries a tyrosine mutation eliminating the p33 termination codon (construct p92Y, Fig. 6A). Since it cannot express p33, construct p92Y alone is not able to replicate in protoplast, confirming that p33 is not expressed (Fig. 6B, lane 20, and Oster et al., 1998). The function of the second RNA, which is based on a DI RNA, is to supply the p33 in the two-component system (Fig. 6A). This RNA, which we named DI-p33 RNA, is not able to replicate in protoplast when present alone (Fig. 6B, lane 21, and Oster et al., 1998). Co-infection of p92Y and DI-p33 RNA into protoplast, however, leads to replication of DI-p33 RNA, while construct p92Y can only replicate poorly (at a level close to the detection limit, Fig. 6B, lane WT, and Oster et al., 1998), but apparently it can supply enough p92 to support the efficient replication of DI-p33 RNA in trans. To test the role of the RPR motif in p33 in tombusvirus replication, we made a series of DI-p33 RNA mutants that included nine of the above RPR mutants (Figs. 2A and 6B). After coelectroporation of N. benthamiana protoplasts with construct p92Y and each of the DI-p33 RNA mutants, we tested the level of accumulation of DI-p33 RNA by Northern blotting (Fig. 6B). The accumulation of three of the DI-p33 RNA mutants was not detected in this assay. These were the mutants with the third arginine being replaced with either alanine or lysine (mutants R646⫺A and R646⫺K, lanes 11, 12, 17 and 18, Fig. 6B) and the combined replacement mutant that carried three alanines in the place of arginines (construct R643– 649⫺A, lanes 13 and 14, Fig. 6B). The DI-p33 RNA mutant with the two prolines within the RPR motif being replaced with alanines (construct P640 – 652⫺A, lanes 7 and 8, Fig. 6B) accumulated poorly (10% of WT level) in protoplast. In contrast, mutants R637⫺A, P652⫺A, P640⫺R, and P640⫺A (lanes 1– 4, 9, 10, 15 and 16, Figs. 6B and C) accumulated to ⬃50% of the WT level, while mutant

198

Z. Panaviene et al. / Virology 308 (2003) 191–205

Z. Panaviene et al. / Virology 308 (2003) 191–205

R649⫺K (lanes 5 and 6, Figs. 6B and C) reached an accumulation level comparable to that of the WT (lanes WT in Figs. 6B and C). Overall, these experiments strongly support that the RPR motif in p33 is essential for replication of tombusviruses. To rule out that the above RPR motif mutants in DI-p33 have a defect in RNA replication due to the altered RNA sequence within the RPR motif, we conducted complementation studies in protoplasts after coelectroporation of the mutated DI-p33 RNA and the full-length gCNV helper, which expressed both WT p33 and WT p92. For the complementation studies, we selected mutants P640 – 652⫺A, R646⫺A, R646⫺K, and R643– 649⫺A, which did not replicate or replicated poorly, and mutant R649⫺K, which replicated efficiently, in the DI-p33/p92Y-based two-component system (Figs. 6B–C). Northern blot analysis of the total RNA extract of the co-inoculated protoplasts revealed that all five DI-p33 mutants tested replicated at levels comparable to that of the WT DI-p33 in the presence of the WT gCNV helper virus after 48 h of incubation in protoplast (compare lanes 1–10 versus lane WT in Fig. 6D). These data demonstrate that the DI-p33 mutants can be complemented in trans by the WT p33 expressed from the gCNV RNA. Therefore, it is unlikely that the RNA genomes of the DI-p33 mutants tested in this work have a cis-acting replication defect due to the altered sequence in the RPR motif. The RPR motif of p92 affects trans replication of DI RNA in protoplast To test the role of the RPR motif in p92 in tombusvirus replication, we also used the p92Y/DI-p33 complementation system (Fig. 7A). For these experiments, we made a series of p92Y RNA mutants (Fig. 7B), similar to the ones tested for p33 in DI-p33 (Fig. 6B). After coelectroporation of N. benthamiana protoplasts with each of the p92Y RPR motif mutants and the WT DI-p33 RNA, we tested the level of accumulation of DI-p33 RNA by Northern blotting (Fig. 7B). These experiments revealed that one of the RPR mutants of p92Y, namely, construct R643– 649⫺A, with three arginine-to-alanine changes (lanes 13 and 14, Figs. 7B and C), did not support the replication of DI-p33 RNA at a detectable level. Another mutant, namely, construct P640 – 652⫺A,

199

with the two prolines being replaced with alanines (lanes 7 and 8, Figs. 7B and C), supported tombusvirus accumulation at a very low level (2.5% of WT). Substitution of alanine for arginine in construct R646⫺A also reduced tombusvirus accumulation by 70% (lanes 11 and 12, Figs. 7B and C). The other six mutants of p92Y tested supported tombusvirus accumulation efficiently (75– 85% of WT, Figs. 7B and C). Overall, the observation that one of the tested p92Y mutants did not support tombusvirus replication confirms that the RPR motif in p92 is essential for tombusvirus replication. The data also suggest that the role of the RPR motif is different in p33 than in p92, since two p33 mutants carrying either alanine or lysine substitutions at position R646 (namely, DI-p33/R646⫺A and R646⫺K) did not support tombusvirus replication at detectable levels (lanes 11, 12, 17 and 18 in Figs. 6B and C), while the corresponding p92Y mutants did support DI-p33 replication (lanes 11, 12, 17 and 18 in Figs. 7B and C).

Discussion The replication process of tombusviruses is currently not yet fully understood. The p92 protein that is expressed from the genomic RNA by a ribosomal readthrough mechanism due to leaky termination at the end of the p33 ORF (Scholthof et al., 1995) contains the signature motifs of viral RdRp within its unique C-terminal region. In contrast, the role of the essential p33 protein in tombusvirus replication is currently unknown. We have recently demonstrated that the C-terminal portion of p33 and the corresponding region of p92 contain an RPR motif that is involved in RNA binding in vitro (Fig. 1, Rajendran and Nagy, submitted for publication). In this paper, we further delineated the role of particular amino acids within the RPR motif in RNA binding in vitro and tombusvirus and DI RNA replication in protoplasts. The in vitro RNA-binding studies revealed that none of the changes within the RPR motif that included one, two, and three amino acid alterations, deletions, or insertions abolished completely the ability of p33 to bind to tombusvirus RNA (Figs. 2A–C). In contrast, deletion of the entire RPR motif did result in an 80% loss in RNA binding (Fig.

Fig. 6. Testing the role of the RPR motif in p33 in tombusvirus replication using a two-component system. (A). Schematic representation of the two constructs used. The stop codon at the end of the p33 gene was mutated to a tyrosine codon (represented by letter Y in the resulting construct, p92Y) in gCNV RNA to prevent the production of p33 from this RNA. The second RNA (DI-p33) was derived from a DI RNA that carried a translation-competent p33 gene (Oster et al., 1998). The RPR motif within the p33, shown with a gray box in DI-p33, was mutated (marked with an asterisk). (B) Northern blot analysis of replication of DI-p33 mutants in the two-component system. The p92Y RNA (4 ␮g) (Oster et al., 1998) was coelectroporated with DI-p33 (2 ␮g) carrying a given mutation within the RPR motif into N. benthamiana protoplasts (5 ⫻ 105 cells per experiment). Sample preparation and Northern blotting were done as described in the legend to Fig. 4, except the probing was done with 32P-labeled RNA specific for DI-p33 RNA (see Materials and Methods). The same amount of total RNA was used for loading onto the gels based on the estimation of the host ribosomal RNA in ethidium bromide-stained gels (see the gel at the bottom of B). The positions of the DI-p33 and p92Y RNAs are depicted on the left. Note that p92Y replicates poorly in protoplasts (undetectable with this probe), while the control gCNV replicates to a high level and the gCNV RNA can be detected with this probe (due to weak cross-hybridization). The names of the DI-p33 mutants are shown on the top. The samples were taken 48 h after electroporation. The experiment was repeated three times. (C) The relative replication efficiency of DI-p33 RNA in the presence of the p92Y helper (WT was chosen as 100%); see legend to Fig. 4 for details. (D) Complementation of replication of DI-p33 mutants by the WT gCNV helper RNA. The Northern blot was performed as described in B.

200

Z. Panaviene et al. / Virology 308 (2003) 191–205

Fig. 7. Testing the role of the RPR motif in p92 in tombusvirus replication using a two-component system. (A) Schematic representation of the two constructs used; see legend to Fig. 6 for further details. The RPR motif within the p92, shown with a gray box, was mutated (marked with an asterisk). (B) Northern blot analysis of replication of WT DI-p33 in the two-component system. The DI-p33 RNA (2 ␮g) was coelectroporated with p92Y RNA (4 ␮g) carrying a given mutation within the RPR motif into N. benthamiana protoplasts (5 ⫻ 105 cells per experiment). Sample preparation and Northern blotting were done as described in the legend to Fig. 4, except the probing was done with 32P-labeled RNA specific for DI-p33 RNA. The names of the p92Y mutants are shown on the top. The samples were taken 48 h after electroporation. The experiment was repeated three times; see Fig. 4 for further details. (C) The relative replication efficiency of WT DI-p33 RNA in the presence of the WT or mutated p92Y helper; see legend to Fig. 4 for details.

1, and Rajendran and Nagy, submitted for publication), suggesting that the entire RPR motif contributes to the strength of RNA binding. Among the individual amino acids in the RPR motif (the sequence is R637P640R643R646R649P652; the third arginine is underlined), the third arginine (R646) and, to a lesser extent, the second arginine (R643) play the most significant roles in RNA binding in vitro. Replacing these

amino acids with alanine resulted in a 40 to 50% reduction in RNA binding in vitro (Figs. 2A–C). Replacing the third arginine with lysine, a positively charged amino acid, also reduced RNA binding by 44% in vitro. Interestingly, similar replacement of either the second or the fourth arginine with lysines increased the efficiency of RNA binding. These data suggest that positively charged amino acids should be present at the

Z. Panaviene et al. / Virology 308 (2003) 191–205

positions of the second and the fourth arginines, while the function of the third arginine (R646) depends not only on its charge, but possibly on its structure as well. Combined deletion of the second, third, and fourth arginines in the RPR motif reduced RNA binding by 62%, but the fact that it did not reduce RNA binding in vitro as much as the deletion of the motif (mutant ⌬RPR, Fig. 1) suggests that the first arginine also plays a role in RNA binding. This is also supported by the observation that alanine substitution for the first arginine (R637) reduced RNA binding by 33%. In contrast to the arginines, the direct role of the two prolines in RNA binding is less significant in vitro. For example, replacing either of the prolines with alanines resulted in RNA binding similar to the WT level (construct P640⫺A and P652⫺A, Figs. 2A–C). On the contrary, replacing both prolines with alanines resulted in a 30% reduction in RNA binding (construct P640 – 652⫺A, Figs. 2A–C). This suggests that the prolines likely play a structural role that may facilitate the correct positioning of arginines for RNA binding. Interestingly, replacing one or both prolines with arginines (constructs P640⫺R, P652⫺R, and P640 – 652⫺R, Figs. 2A–C) increased RNA binding by 2.5- to 3-fold, indicating that the WT RPR motif is not the most efficient sequence in RNA binding. Importantly, however, these mutants supported cis or trans replication inefficiently (Figs. 4 and 5A and B), suggesting that the ability of p33 and/or p92 to bind strongly to RNA is actually detrimental for replication. Detailed testing of the RPR motif mutants in protoplasts revealed that this motif is essential for tombusvirus replication. For example, 10 of the 19 RPR mutants tested replicated poorly in protoplast (Figs. 4A and B). These mutants also supported trans replication of DI-72 RNA poorly (Figs. 5A and B), suggesting that the reduction in the level of tombusvirus RNA replication is likely due to the mutations that render p33 and/or p92 trans-acting factors less efficient. The use of a complementation-based two-component system, which included the tombusvirus genomic RNA expressing only p92 (but not the p33 protein due to the elimination of the leaky stop codon in the p33 ORF) and a DI RNA producing p33, defined that mutations within the RPR motif of both p33 and p92 can debilitate tombusvirus replication. These data (Figs. 6 and 7) confirmed that the RPR motif, and therefore likely the ability to bind RNA, is important for the function of both p33 and p92 during tombusvirus infection. Interestingly, however, we observed significant differences among particular mutations depending on whether they were located in p33 or p92. For example, mutating the third arginine in the RPR motif in p33 to alanine or lysine (constructs R646⫺A and R646⫺K) reduced tombusvirus replication below the level of detection. The same mutations when present in p92, however, only reduced the level of RNA accumulation from 30 to 90% of the WT level (Figs. 6 and 7A and B). The ability of the same mutants of p33 and p92 to support tombusvirus replication at different levels suggests that the function/structure of the RPR motif is different when present in p33 or p92. It is also

201

possible that the above differences may reflect quantitative differences between p33 and p92 during infection. Indeed, p33 is expressed approximately at 20-fold higher levels in plants than p92 (Scholthof et al., 1995), suggesting that more p33 may be needed for replication than p92. Therefore, reducing the ability of p33 to bind RNA may affect replication more than that of p92, which has a second RNA-binding region (Rajendran and Nagy, submitted for publication). Alternatively, the RPR motif may overlap with another, yet unknown, domain in either p33 or p92 that is also involved in tombusvirus replication. If this is the case, then mutations within the RPR motif may also affect the function of the second putative domain in either p33 or p92, resulting in differences in RNA replication in the twocomponent system. Comparison of the ability of p33 to bind RNA in vitro and the efficiency of tombusvirus replication in protoplasts revealed three groups of mutants. The first group includes seven RPR motif mutants that bound to RNA by less than 60% of the WT level in vitro. Six of these mutants supported a dramatically decreased level of gCNV accumulation (10% or less than the WT level) in protoplasts (see constructs R646⫺A, R646⫺K, ⌬R, ⌬2R, ⌬3R, and R643– 649⫺A; Figs. 2, 4 and 5). These observations are valid for both cis replication (i.e., for gCNV RNA accumulation) and trans replication (i.e., for DI-72 RNA accumulation). The only somewhat unusual mutant in this group is R643⫺A, which supported gCNV accumulation at 20% of the WT level, but DI-72 RNA was replicated by over 80% of the WT level. Overall, the results obtained with this group of mutants suggest that the ability to bind to RNA efficiently (more than 60% of the WT level) may be important for the function of p33 and/or p92. The second group of mutants includes four members (R637⫺A, P652⫺A, P640 – 652⫺A, and R649⫺A) that are similar to the first group (i) in their ability to bind RNA less efficiently in vitro than the WT (between 5 and 30% reduction) and (ii) they also support tombusvirus replication less efficiently than WT (between 5 and 60% reduction), but they are different from the first group because the effects on RNA binding and replication are not as dramatic as in the case of the first group. The third group of mutants includes eight members that are able to bind RNA more efficiently than the WT protein (an increase between 10 to 200%, Fig. 2). Five of these mutants support tombusvirus replication at a very low level (10% or less, see mutants P652⫺R, P640 – 652⫺R, R643⫺K, R643⫹R, and R643⫹2R, Figs. 4 and 5). The remaining three of these mutants also show a reduced level of replication, but the reduction is less pronounced (between 25 and 50%, see constructs P640⫺A, P640⫺R, and R649⫺K). Overall, these results suggest that the ability to bind RNA more strongly than the WT, like the ability to bind weakly, is also detrimental for tombusvirus replication. This suggests that the WT p33 and p92 replicase proteins of tombusviruses contain the sequence that may be optimal for both the binding and the release of RNA templates during replica-

202

Z. Panaviene et al. / Virology 308 (2003) 191–205

tion. In contrast, it is possible that those mutants that can bind RNA either too weekly or too strongly are deficient in replication. Comparison of the RPR mutants for their abilities to support viral replication revealed that the third arginine (R646) and, to a lesser extent, the second arginine (R643) affected tombusvirus replication to the largest extent among the single alanine mutants. Importantly, these mutations also inhibited in vitro RNA binding to the highest extent (Fig. 2). Replacement of the first and the fourth arginines with alanines affected tombusvirus replication and RNA binding to a lesser extent than that of the second and the third arginines did. Deletion of one, two, or three of the arginines in the RPR motif reduced cis replication below the level of detection, while the accumulation of the DI RNA (in the trans replication system) was also reduced to 0 to 1% of the WT level in vivo. Similarly, replacing three of the arginines with alanines resulted in tombusvirus accumulation below the level of detection. These data confirm that the arginines of the RPR motif are essential for tombusvirus replication. In contrast to the arginine mutations, replacement of the first or the second or both prolines with alanines affected the level of tombusvirus accumulation only by 30 to 60% in protoplasts. Overall, the fact that several alanine mutants can support tombusvirus replication rather efficiently suggests that p33 and/or p92 may tolerate some loss in effectiveness of RNA binding without major detrimental effects on their functions under the conditions used in vivo. The possibility may exist that these mutants may have lost fitness that would only be obvious under competitive conditions (e. g., in the presence of WT p33/p92). The interpretation of in vitro RNA binding data and the in vivo RNA replication data obtained with p33/p92 mutants carrying arginine or lysine substitutions or insertions is certainly more complex than that with the corresponding alanine mutants. For example, separate replacement of the second and third arginines with lysines resulted in 130 and 55% RNA binding, respectively, while both mutants replicated only at ⬃5% of the level of the WT. Lysine substitution for the fourth arginine in the RPR motif increased RNA binding to 285%, and it reduced the level of replication to 50% when compared to WT. Similar to the above lysine mutants, the interpretation of data obtained with the proline to arginine mutants and for mutants carrying extra arginines within the RPR motif is not straightforward. For example, addition of one or two extra arginines increased the ability of p33 to bind RNA 20 to 30% more efficiently, but it inhibited replication dramatically (down to 0 to 10% of WT). In addition, our data show that the two prolines in the RPR motif can be replaced with small neutral amino acids (such as alanine), but not with the large positively charged arginine(s) without debilitating cis or trans replication of tombusviruses. To explain all these data, we propose that for the proper functions of p33 and p92, not only may it be important to bind RNA, but it is also critical to form the “right” structure with the viral RNA. For example,

it is possible that the RPR motif in p33 and/or p92 is needed for proper presentation of the RNA in the viral replicase complex during replication. Mutations within the RPR motif might change the positioning of the RNA within p33 and/or p92 that could affect the efficiency of replication. In addition to the proper presentation of the RNA, it is also likely that the release of the RNA after replication is important and those features of the p33/p92 proteins are expected to affect the in vivo replication results but these features have not been measured with the gel mobility-shift assay. Overall, the observed differences between the in vitro RNA-binding and the in vivo replication results may reflect the multiple functions of the RNA binding domain in p33/ p92, such as binding, positioning, and releasing the RNA, all of which might be critical for RNA replication in the infected cells. Interestingly, most of the RPR mutants tested supported cis replication of gCNV and trans replication of DI-72 RNA at somewhat similar relative levels when compared to WT gCNV and WT DI-72 RNA. The two notable exceptions are (i) P640⫺R, which supported gCNV replication efficiently (60% of WT, Fig. 4B) and DI-72 RNA replication inefficiently (20% of WT, Fig. 5B), and (ii) R643⫺A, which supported gCNV replication inefficiently (20% of WT), while it supported DI-72 RNA replication at 85% of the WT level (Fig. 5B). Although the reason for the above differences is not yet known, they do suggest that the function(s) of the RPR motif in p33 and/or p92 is not completely the same in cis versus trans replication. Another interesting finding in this work is the altered ratio between genomic and subgenomic RNAs for several mutants when compared to WT. For example, the ratio between sg2 and gCNV was ⬃twentyfold higher in the case of R643⫹R than for WT CNV, while P640 – 652⫺A generated sg2 RNA 40% less efficiently than did WT CNV (Fig. 4C). The currently popular model of subgenomic RNA synthesis for tombusviruses predicts that the viral replicase terminates prematurely during the minus-strand synthesis in the vicinity of the subgenomic promoter (Choi et al., 2001). The truncated minus-stranded RNA then serves as a template for the synthesis of plus-stranded subgenomic RNAs from the subgenomic promoter. It is possible that the above RPR mutants may have either higher (for example, R643⫹R) or lower (P640 – 652⫺A) termination rates in the vicinity of the subgenomic promoter than WT, which then could lead to altered sg2 and, to a lesser extent, sg1 RNA synthesis. Further in vitro experiments with purified p33/p92 mutants will be needed to address this question.

Materials and methods Site-directed mutagenesis of CNV (pK2M5) A full-length cDNA clone of CNV, named pK2M5 (Rochon and Johnston, 1991), was used to generate the RPR

Z. Panaviene et al. / Virology 308 (2003) 191–205

203

Table 1 List of primer used for site-directed mutagenesis of CNV(pK2M5) Mutant

Primer pair

Sequencea

Positionb

R637⫺A

554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 594 595 596 597

ATTGCGTCCACAGGAGCGCCTCGCAGAAGACC AGGTCTTCTGCGAGGCGCTCCTGTGGACGCAAT GCGTCCACAGGACGCGCGCGCAGAAGACCTTATG CATAAGGTCTTCTGCGCGCGCGTCCTGTGGACGC TCCACAGGACGCCCTGCGAGAAGACCTTATGCA TGCATAAGGTCTTCTCGCAGGGCGTCCTGTGGA ACAGGACGCCCTCGCGCGAGACCTTATGCAGCT AGCTGCATAAGGTCTCGCGCGAGGGCGTCCTGT GGACGCCCTCGCAGAGCGCCTTATGCAGCTAAG CTTAGCTGCATAAGGCGCTCTGCGAGGGCGTCC CGCCCTCGCAGAAGAGCGTATGCAGCTAAGATTG CAATCTTAGCTGCATACGCTCTTCTGCGAGGGCG TCCACAGGACGCCCTAAGAGAAGACCTTATGCA TGCATAAGGTCTTCTCTTAGGGCGTCCTGTGGA ACAGGACGCCCTCGCAAGAGACCTTATGCAGCT AGCTGCATAAGGTCTCTTGCGAGGGCGTCCTGT GGACGCCCTCGCAGAAAGCCTTATGCAGCTAAG CTTAGCTGCATAAGGCTTTCTGCGAGGGCGTCC TCCACAGGACGCCCTAGACGCAGAAGACCTTATG CATAAGGTCTTCTGCGTCTAGGGCGTCCTGTGGA TCCACAGGACGCCCTAGAAGACGCAGAAGACCTTATG CATAAGGTCTTCTGCGTCTTCTAGGGCGTCCTGTGGA TCCACAGGACGCCCT–AGAAGACCTTATGCA TGCATAAGG–TCTTCTAGGGCGTCCTGTGGA TCCACAGGACGCCCT––AGACCTTATGCAGCT AGCTGCATAAGG––TCTAGGGCGTCCTGTGGA TCCACAGGACGCCCT–––CCTTATGCAGCTAAG CTTAGCTGCATAAGG–––AGGGCGTCCTGTGGA GCGTCCACAGGACGCGCGCGCAGAAGAGCGTATGCAGCTAAGATTG CAATCTTAGCTGCATACGCTCTTCTGCGCGCGCGTCCTGTGGACGC GCGTCCACAGGACGCAGACGCAGAAGAAGATATGCAGCTAAGATTG CAATCTTAGCTGCATATCTTCTTCTGCGTCTGCGTCCTGTGGACGC TCCACAGGACGCCCTGCGGCGGCGCCTTATGCAGCTAAG CTTAGCTGCATAAGGCGCCGCCGCAGGGCGTCCTGTGGA GCGTCCACAGGACGCAGACGCAGAAGACCTTATG CATAAGGTCTTCTGCGTCTGCGTCCTGTGGACGC CGCCCTCGCAGAAGAAGATATGCAGCTAAGATTG CAATCTTAGCTGCATATCTTCTTCTGCGAGGGCG

622–653 654–622 625–658 658–625 628–660 660–628 631–663 663–631 634–666 666–634 637–670 670–637 628–660 660–628 631–663 663–631 634–666 666–634 628–658 658–628 628–658 658–628 628–660 660–628 628–663 663–628 628–666 666–628 625–670 670–625 625–670 670–625 628–666 666–628 625–658 658–625 637–670 670–637

P640⫺A R643⫺A R646⫺A R649⫺A P652⫺A R643⫺K R646⫺K R649⫺K R643⫹R R643⫹2R ⌬R ⌬2R ⌬3R P640–652⫺A P640–652⫺R R643–649⫺A P640⫺R P652⫺R a b

The mutated sequences are underlined. The nucleotide positions within the p92ORF are indicated (Rochon and Johnston, 1991).

motif mutants of p33/p92 listed in Fig. 2A. Mutagenesis was performed by PCR using the QuickChange XL SiteDirected Mutagenesis Kit (Stratagene). The PCR reactions included the Pfu Turbo DNA polymerase, different sets of primer pairs designed for each mutant (see Table 1), and 40 ng of pK2M5 DNA as a template. PCR products were digested with DpnI before transformation into E. coli (DH5␣). The presence of the desired mutations was confirmed by sequencing with primer 27 (5⬘-GTATTTCACACCAAGGGAC-3⬘). The series of p92Y constructs, containing mutated or WT RPR motif (Fig. 7), was obtained by replacing the EagI and NsiI (positions 998 and 1650 in the p92 ORF) fragment with the corresponding mutated fragment of Mal/p92 (Rajendran and Nagy, submitted for publication), containing a stop codon to tyrosine codon mutation at the end of the p33 gene. The presence of the tyrosine mutation was confirmed in

each clone by sequencing, using primer 631 (5⬘-GAGGAATTCAAGGTAATTGCGTCCACA-3⬘). For the construction of DI-p33 series of mutants (Fig. 6), we used construct DI-83CNV (a generous gift of Andy White) (Oster et al., 1998). Each full-length RPR-motif mutant of CNV was digested with StuI and SphI and the DNA fragment containing the p33 sequence was used to replace the corresponding sequence in DI-83CNV. For the gel shift experiments, we used either the full length or a truncated version of the p33 protein that was fused to MBP (vector pMal-c2X, New England Biolabs). To make the p33 expression constructs, containing the WT or each of the mutated RPR motifs shown in Fig. 2A, we amplified a 90-nt-long fragment of the p33 gene (positions 616 –705 in the p92 ORF) that represent a 30-amino-acidlong region including the RPR motif by PCR using primers 631 (5⬘-GAGGAATTCAAGGTAATTGCGTCCACA-3⬘)

204

Z. Panaviene et al. / Virology 308 (2003) 191–205

and 632 (5⬘-GAGTCTAGACTACTTCAGGTAACCCACCTT-3⬘) with Deep Vent polymerase (New England Biolabs). The templates for the PCR were WT or mutated pK2M5 DNA (see above). The PCR fragments were digested with EcoRI and XbaI, gel isolated, and ligated into the similarly treated pMal-c2X vector. To generate constructs expressing full-length p33 fused with MBP (Figs. 1 and 3), the WT or mutant p33 sequences were PCR amplified with primers 25 (5⬘-GGA GTC TAG AGA TAC CAT CAA GAG GAT G-3⬘) and 992B (5⬘GAGCTGCAGCTATTTCACACCAAGGGA-3⬘) using Deep Vent polymerase (see above). The PCR products were digested with XbaI and PstI and cloned into the pMAL-c2 vector to generate in-frame MBP-p33 fusion. To obtain the expression construct ⌬RPR (Fig. 1), we amplified the 5⬘ and 3⬘ portions of the CNV p33 gene separately using either primer pair 25 (above) and 994B (5⬘-CCGCGCTAGCTCCTGTGGACGCAATTACCT-3⬘) or 993B (5⬘-GCGGGCTAGCTATGCAGCTAAGATTGCACA-3⬘) and 992 (above). The PCR product representing the 5⬘ portion of the p33 gene was digested with XbaI and NheI, while the PCR product representing the 3⬘ portion of the p33 gene was digested with NheI and PstI. Both PCR products were cloned into pMAL-c2, which had been digested with PstI and XbaI, in a three-piece ligation reaction. The clones were sequenced to validate their correctness. Preparation of RNA transcripts RNA transcripts were obtained using SmaI-linearized WT pK2M5, mutated pK2M5, DI-72, and DI-p33 (the original name is DI-83CNV, Oster et al., 1998) clones in a standard transcription reaction with T7 RNA polymerase (Nagy et al., 1999). Template DNA was removed by DNase I, followed by purification of the RNA transcript with phenol/chloroform extraction and 95% ethanol precipitation. The pellet was washed three times with 70% ethanol to remove residual salts. The RNA transcripts were quantified by UV spectrophotometer (Beckman), followed by 1% agarose gel electrophoresis. Purification of full-length and truncated p33 from E. coli The protein expression for each p33 RPR-motif mutant (Fig. 2A) was induced at 37°C by IPTG in Epicurian BL21Codon plus (DE3)-RIL cells (Strategene) as recommended by the supplier. After 2 h of induction, the cells were harvested and processed as described by Rajendran et al. (2002). The proteins were eluted from the amylose column (New England Biolabs) according to Rajendran et al. (2002). All steps were carried out on ice or in the cold room. The quality of the proteins obtained was checked by 10% SDS–PAGE analysis and their amounts were measured as described by Rajendran et al. (2002). The RNA-binding studies (see below) were done with the fusion proteins.

In vitro RNA-binding studies The RNA probe in the gel shift experiments was the 169-nt region I (⫹) of DI-72 (White and Morris, 1994), which was labeled with [32P]UTP using T7 RNA polymerase (Rajendran et al., 2002). The gel-shift experiments were performed according to Rajendran et al. (2002). Briefly, ⬃2 ng of [32P]UTP-labeled RNA was mixed with 1 ␮g of the p33 preparation for 15 min at 25°C in the presence of 50 mM Tris–HCl, pH 8.2, 10 mM MgCl2, 10 mM DTT, 10% glycerol, 2.4 U of RNase inhibitor, and 100 ng of tRNA. The samples were analyzed by electrophoresis on native 4% polyacrylamide gels run at 200 V for 50 min at 4°C in TAE buffer. Dried gels were analyzed using a phosphorimager. Preparation of protoplasts N. benthamiana protoplasts were prepared from callus cells by treatment with 0.5 g of cellulysin and 0.1 g of macerase (Calbiochem) in 50 ml of protoplast incubation media (Kong et al., 1997; Nagy et al., 1999, 2001) for 4.5 h with gentle shaking at 25°C. After incubation, the protoplasts were filtrated through a sieve set and a sterile nylon cloth. The protoplasts were washed twice with 0.5 M mannitol and once with the electroporation buffer (10 mM HEPES, 10 mM NaCl, 120 mM KCl, 4 mM CaCl2, 200 mM mannitol); 5 ⫻ 105 protoplasts were electroporated with either 2 ␮g of in vitro transcribed gCNV RNA or the combination of 2 ␮g of gCNV RNA and 1 ␮g of DI-72 RNA. In the two-component studies (Figs. 6 and 7), we used 4 ␮g of p92Y and 2 ␮g of DI-p33 RNA transcripts. Electroporation was done using Gene pulser II (Bio–Rad, the settings were at 0.2 kV and 0.5 ␮F). After electroporation, samples were kept on ice for 30 min, followed by adding 1.8 ml of protoplast culture medium (Kong et al., 1997) to each sample. Protoplasts were incubated in 35 ⫻ 10 mm petri dishes in the dark for 24 – 48 h at 22°C. Total RNA extraction from protoplasts and RNA analysis Total RNA was extracted using the phenol/chloroformbased method of Kong et al. (1997) and Nagy et al. (2001). For Northern blot analysis, the total RNA was treated with formamide at 85°C, followed by agarose gel electrophoresis (0.8% for gCNV RNA and 1.2% for DI RNA). After electrophoresis, the RNA was transferred by electrotransfer to Hybond XL membrane (Amersham-Pharmacia) and hybridized with gCNV- or DI-72-specific probes in ULTRAhyb hybridization buffer at 68°C using the conditions recommended by the supplier (Ambion). The probes were prepared by T7 RNA polymerase in an in vitro transcription reaction in the presence of [␣-32P]UTP using PCR-amplified DNA templates that were obtained by using primers 16 (5⬘-GTAATACGACTCACTATAGGGCTGCATTTCTGCAATGTTC-3⬘) and 312 (5⬘-GCTGTCAGTCTAGTGGA3⬘) for the gCNV and primers 15 (5⬘-GTAATA-

Z. Panaviene et al. / Virology 308 (2003) 191–205

CGACTCACTATAGGGCATGTCGCTTGTTTGTTG-3⬘) and 20 (5⬘-GGAAATTCTCCAGGATTTCTC-3⬘) for DI-72 RNA. These probes are complementary to the 3⬘-terminal 200 nt of gCNV RNA and the 5⬘ 169 nt of DI-72 RNA, respectively. RT–PCR analysis of gCNV RNA from protoplasts We tested the stability of the RPR motif mutants of gCNV during virus replication in protoplasts by sequencing the RT–PCR products obtained from total RNA samples using primer 27 (5⬘-GTATTTCACACCAAGGGAC-3⬘) for RT and primers 213 (5⬘-GAGGATGAAAGGGAATTCACGGATTGTTTGG-3⬘) and 27 for the PCR reaction. The RT–PCR products were gel purified before sequencing, which was performed by using primer 27 (see above).

Acknowledgments We thank Dr. Judit Pogany and Dr. Tadas Panavas for critical reading of the manuscript and for very helpful suggestions. We are grateful to Dr. Andy White for providing the DI83CNV clone and for helpful suggestions. This work was supported by NSF Grant MCB0078152 and by the Kentucky Tobacco Research and Development Center at the University of Kentucky. This study is Publication No. 0212-123 of the Kentucky Agricultural Experiment Station.

References Ahlquist, P., 2002. RNA-dependent RNA polymerases, viruses, and RNA silencing. Science 296 (5571), 1270 –1273. Buck, K.W., 1996. Comparison of the replication of positive-stranded RNA viruses of plants and animals. Adv. Virus Res. 47, 159 –251. Buck, K.W., 1999. Replication of tobacco mosaic virus RNA. Philos. Trans. R. Soc. London Ser. B Biol. Sci. 354 (1383), 613– 627. Burgyan, J., Rubino, L., Russo, M., 1991. De novo generation of cymbidium ringspot virus defective interfering RNA. J. Gen. Virol. 72 (Pt. 3), 505–509. Chang, Y.C., Borja, M., Scholthof, H.B., Jackson, A.O., Morris, T.J., 1995. Host effects and sequences essential for accumulation of defective interfering RNAs of cucumber necrosis and tomato bushy stunt tombusviruses. Virology 210 (1), 41–53. Choi, I.R., Ostrovsky, M., Zhang, G., White, K.A., 2001. Regulatory activity of distal and core RNA elements in Tombusvirus subgenomic mRNA2 transcription. J. Biol. Chem. 276 (45), 41761– 41768. Diez, J., Ishikawa, M., Kaido, M., Ahlquist, P., 2000. Identification and characterization of a host protein required for efficient template selection in viral RNA replication. Proc. Natl. Acad. Sci. USA 97 (8), 3913–3918.

205

Finnen, R.L., Rochon, D.M., 1993. Sequence and structure of defective interfering RNAs associated with cucumber necrosis virus infections. J. Gen. Virol. 74 (Pt. 8), 1715–1720. Hillman, B.I., Carrington, J.C., Morris, T.J., 1987. A defective interfering RNA that contains a mosaic of a plant virus genome. Cell 51 (3), 427– 433. Knorr, D.A., Mullin, R.H., Hearne, P.Q., Morris, T.J., 1991. De novo generation of defective interfering RNAs of tomato bushy stunt virus by high multiplicity passage. Virology 181 (1), 193–202. Kollar, A., Burgyan, J., 1994. Evidence that ORF 1 and 2 are the only virus-encoded replicase genes of cymbidium ringspot tombusvirus. Virology 201 (1), 169 –172. Kong, Q., Wang, J., Simon, A.E., 1997. Satellite RNA-mediated resistance to turnip crinkle virus in Arabidopsis involves a reduction in virus movement. Plant Cell 9 (11), 2051–2063. Lai, M.M., 1998. Cellular factors in the transcription and replication of viral RNA genomes: a parallel to DNA-dependent RNA transcription. Virology 244 (1), 1–12. Nagy, P.D., Pogany, J., 2000. Partial purification and characterization of Cucumber necrosis virus and Tomato bushy stunt virus RNA-dependent RNA polymerases: similarities and differences in template usage between tombusvirus and carmovirus RNA-dependent RNA polymerases. Virology 276 (2), 279 –288. Nagy, P.D., Pogany, J., Simon, A.E., 1999. RNA elements required for RNA recombination function as replication enhancers in vitro and in vivo in a plus-strand RNA virus. EMBO J. 18 (20), 5653–5665. Nagy, P.D., Pogany, J., Simon, A.E., 2001. In vivo and in vitro characterization of an RNA replication enhancer in a satellite RNA associated with turnip crinkle virus. Virology 288 (2), 315–324. Noueiry, A.O., Chen, J., Ahlquist, P., 2000. A mutant allele of essential, general translation initiation factor DED1 selectively inhibits translation of a viral mRNA. Proc. Natl. Acad. Sci. USA 97 (24), 12985– 12990. Oster, S.K., Wu, B., White, K.A., 1998. Uncoupled expression of p33 and p92 permits amplification of tomato bushy stunt virus RNAs. J. Virol. 72 (7), 5845–5851. Panavas, T., Pogany, J., Nagy, P.D., 2002. Analysis of minimal promoter sequences for plus-strand synthesis by the cucumber necrosis virus RNA-dependent RNA polymerase. Virology 296 (2), 263–274. Quadt, R., Kao, C.C., Browning, K.S., Hershberger, R.P., Ahlquist, P., 1993. Characterization of a host protein associated with brome mosaic virus RNA-dependent RNA polymerase. Proc. Natl. Acad. Sci. USA 90 (4), 1498 –1502. Rajendran, K.S., Pogany, J., Nagy, P.D., 2002. Comparison of turnip crinkle virus RNA-dependent RNA polymerase preparations expressed in Escherichia coli or derived from infected plants. J. Virol. 76 (4), 1707–1717. Rochon, D.M., Johnston, J.C., 1991. Infectious transcripts from cloned cucumber necrosis virus cDNA: evidence for a bifunctional subgenomic mRNA. Virology 181 (2), 656 – 665. Scholthof, K.B., Scholthof, H.B., Jackson, A.O., 1995. The effect of defective interfering RNAs on the accumulation of tomato bushy stunt virus proteins and implications for disease attenuation. Virology 211 (1), 324 –328. White, K.A., Morris, T.J., 1994. Recombination between defective tombusvirus RNAs generates functional hybrid genomes. Proc. Natl. Acad. Sci. USA 91 (9), 3642–3646. White, K.A., Morris, T.J., 1999. Defective and defective interfering RNAs of monopartite plus-strand RNA plant viruses. Curr. Top. Microbiol. Immunol. 239, 1–17.